INDEX

Explanations

HTML or document formatting elements related to font size and capitalization

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

']").

-0.48

 giudi

-0.48

 }}"></

-0.47

SEGUIR

-0.47

inSlope

-0.46

 estekak

-0.46

}`).

-0.46

bukkit

-0.44

 Suivez

-0.43

 seyn

-0.43

POSITIVE LOGITS

uppercase

1.15

lowercase

0.80

 uppercase

0.75

 Exactos

0.60

 lowercase

0.60

toUpperCase

0.57

 '\\;'

0.55

 AssemblyTitle

0.52

WriteTagHelper

0.51

unsplash

0.50

Activations Density 0.013%

HTML or document formatting elements related to font size and capitalization

No Comments

No Known Activations